On 01/03/2013 06:59 PM, dill...@dillona.com wrote:
> From: Dillon Amburgey <dill...@dillona.com>
>
> This is used at least on the Alpha architecture
>
> Signed-off-by: Dillon Amburgey <dill...@dillona.com>
> ---
> linux-user/syscall.c | 8 ++++++++
> 1 files changed, 8 insertions(+), 0 deletions(-)
>
> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> index e99adab..0cd54f3 100644
> --- a/linux-user/syscall.c
> +++ b/linux-user/syscall.c
> @@ -5497,6 +5497,14 @@ abi_long do_syscall(void *cpu_env, int num, abi_long
> arg1,
> unlock_user(p3, arg3, 0);
> break;
> }
> +#ifdef TARGET_NR_oldumount
> + case TARGET_NR_oldumount:
> + if (!(p = lock_user_string(arg1)))
> + goto efault;
> + ret = get_errno(umount(p));
> + unlock_user(p, arg1, 0);
> + break;
> +#endif
It's probably better to rename the Alpha TARGET_NR_{oldumount,umount}
syscall names to TARGET_NR_{umount,umount2}. Regardless of what the
names the kernel uses for these, it's the mapping to the implementations
in the syscall tables that matters.
r~
